Acme Support
demo
A support agent that regressed when someone changed its model.
This is a worked example
Nothing here came from your code — it is a support agent that regressed when somebody changed its model. Everything works, so open a run and follow it down to the rows that broke. Read-only, and safe to delete whenever it stops being useful.
| Suite | Latest score | Pass rate | Runs | Last run |
|---|---|---|---|---|
|
SupportBotQuality
App\Evals\SupportBotQuality
|
83.1% ▼12.0 | 75.0% | 7 | 4 days ago |
|
TicketTriageStructure
App\Evals\TicketTriageStructure
|
100.0% | 100.0% | 4 | 5 days ago |
|
RefundPolicyAccuracy
App\Evals\RefundPolicyAccuracy
|
87.1% ▲0.9 | 83.0% | 3 | 1 week ago |